Arguments#

EXPERIMENT-RELATED ARGS

--dataset<class ‘str’>

Help: Which dataset to perform experiments on.

  • Default: None

  • Choices: seq-tinyimg, seq-tinyimg-r, perm-mnist, seq-cifar10, seq-cifar100-224, seq-cub200, rot-mnist, seq-cifar100, seq-cifar100-224-rs, seq-mnist, mnist-360

--model<function custom_str_underscore at 0x7fe9f47a42c0>

Help: Model name.

  • Default: None

  • Choices: agem, agem-r, ewc-on, derpp-lider, gdumb-lider, slca, dualprompt, si, bic, er-ace, fdr, gdumb, gem, gss, joint-gcl, lwf, mer, rpc, twf, ccic, der, derpp, er, hal, icarl, l2p, lucir, lwf-mc, sgd, xder, xder-ce, xder-rpc, pnn, er-ace-lider, icarl-lider, coda-prompt

--lr<class ‘float’>

Help: Learning rate.

  • Default: None

  • Choices:

--optimizer<class ‘str’>

Help: Optimizer.

  • Default: sgd

  • Choices: sgd, adam, adamw

--optim_wd<class ‘float’>

Help: optimizer weight decay.

  • Default: 0.0

  • Choices:

--optim_mom<class ‘float’>

Help: optimizer momentum.

  • Default: 0.0

  • Choices:

--optim_nesterov<class ‘int’>

Help: optimizer nesterov momentum.

  • Default: 0

  • Choices:

--lr_scheduler<class ‘str’>

Help: Learning rate scheduler.

  • Default: None

  • Choices:

--lr_milestones<class ‘int’>

Help: Learning rate scheduler milestones (used if lr_scheduler=multisteplr).

  • Default: []

  • Choices:

--sched_multistep_lr_gamma<class ‘float’>

Help: Learning rate scheduler gamma (used if lr_scheduler=multisteplr).

  • Default: 0.1

  • Choices:

--n_epochs<class ‘int’>

Help: Number of epochs.

  • Default: None

  • Choices:

--batch_size<class ‘int’>

Help: Batch size.

  • Default: None

  • Choices:

--distributed<class ‘str’>

Help: Enable distributed training?

  • Default: no

  • Choices: no, dp, ddp

--savecheckNone

Help: Save checkpoint?

  • Default: False

  • Choices:

--loadcheck<class ‘str’>

Help: Path of the checkpoint to load (.pt file for the specific task)

  • Default: None

  • Choices:

--ckpt_name<class ‘str’>

Help: (optional) checkpoint save name.

  • Default: None

  • Choices:

--start_from<class ‘int’>

Help: Task to start from

  • Default: None

  • Choices:

--stop_after<class ‘int’>

Help: Task limit

  • Default: None

  • Choices:

--joint<class ‘int’>

Help: Train model on Joint (single task)?

  • Default: 0

  • Choices: 0, 1

--label_perc<class ‘float’>

Help: Percentage in (0-1] of labeled examples per task.

  • Default: 1

  • Choices:

MANAGEMENT ARGS

--seed<class ‘int’>

Help: The random seed.

  • Default: None

  • Choices:

--permute_classes<class ‘int’>

Help: Permute classes before splitting tasks (applies seed before permute if seed is present)?

  • Default: 0

  • Choices: 0, 1

--base_path<class ‘str’>

Help: The base path where to save datasets, logs, results.

  • Default: ./data/

  • Choices:

--notes<class ‘str’>

Help: Notes for this run.

  • Default: None

  • Choices:

--non_verbose<class ‘int’>

Help: Make progress bars non verbose

  • Default: 0

  • Choices: 0, 1

--disable_log<class ‘int’>

Help: Disable logging?

  • Default: 0

  • Choices: 0, 1

--num_workers<class ‘int’>

Help: Number of workers for the dataloaders (default=infer from number of cpus).

  • Default: None

  • Choices:

--validation<class ‘int’>

Help: Percentage of validation set drawn from the training set.

  • Default: None

  • Choices:

--enable_other_metrics<class ‘int’>

Help: Enable computing additional metrics: forward and backward transfer.

  • Default: 0

  • Choices: 0, 1

--debug_mode<class ‘int’>

Help: Run only a few forward steps per epoch

  • Default: 0

  • Choices: 0, 1

--wandb_entity<class ‘str’>

Help: Wandb entity

  • Default: None

  • Choices:

--wandb_project<class ‘str’>

Help: Wandb project name

  • Default: mammoth

  • Choices:

--eval_epochs<class ‘int’>

Help: Perform inference intra-task at every eval_epochs.

  • Default: None

  • Choices:

--inference_onlyNone

Help: Perform inference only for each task (no training).

  • Default: False

  • Choices:

REEHARSAL-ONLY ARGS

--buffer_size<class ‘int’>

Help: The size of the memory buffer.

  • Default: None

  • Choices:

--minibatch_size<class ‘int’>

Help: The batch size of the memory buffer.

  • Default: None

  • Choices:

Functions

utils.args.add_experiment_args(parser)[source]#

Adds the arguments used by all the models.

Parameters:

parser (ArgumentParser) – the parser instance

Returns:

None

Return type:

None

utils.args.add_management_args(parser)[source]#

Adds the management arguments.

Parameters:

parser (ArgumentParser) – the parser instance

Returns:

None

Return type:

None

utils.args.add_rehearsal_args(parser)[source]#

Adds the arguments used by all the rehearsal-based methods

Parameters:

parser (ArgumentParser) – the parser instance

Returns:

None

Return type:

None